CN102110294B - Method and system for processing image of diseased fish body - Google Patents

Abstract

The invention provides a method and a system for processing the image of a diseased fish body. The method for processing the image of a diseased fish body comprises the following steps: obtaining the image of the diseased fish body; obtaining the image of a fish body zone from the fish body image; obtaining an abnormal target zone image from the image of a fish body zone; and removing a non-speck target zone image from the abnormal target zone image to obtain a speck target image. According to the method and the system for processing the image of a diseased fish body, the corresponding speck target image of the speck of the diseased fish can be quickly and accurately obtained to provide basis data for automatically recognizing fish diseases by a computer, so that a technician can conveniently know the specific situation of the diseases of the diseased fish to take measures in time so as to prevent the fulminating disease of aquaculture fish from being spread, thereby lowering loss brought by the fulminating disease of aquaculture fish. In addition, the method for processing the image of a diseased fish body is simple to operate and has good adaptability.

Description

Fish volume image disposal route and the system of sick fish

Technical field

The present invention relates to image processing techniques, relate in particular to fish volume image disposal route and the system of a kind of sick fish.

Background technology

At present, China fish culture industry just develops towards the direction that intensive style is cultured, and the breed scale constantly enlarges, and along with the developing rapidly of fish culture industry, the probability that the aquaculture fish suffer from explosive disease constantly increases.The explosive disease of aquaculture fish is a kind of acute comprehensive fish disease, and it has disease and plants variation, and area is extensively changed, and the big characteristics of treatment difficulty bring very big loss can for the fish culture family.

Because the explosive disease of China aquaculture fish mainly is present in chinese carp at present, promptly in black carp, grass carp, silver carp, bighead, carp and the crucian, the body surface color of the fish body of ill sick fish is different fully with the fish body of normal fish with texture.For example, with regard to bream, it is shiny black that the fish surface of normal bream generally has the back, partly turns white the characteristics of clean mark and shape intact near fish maw; And when ill, the red trace of blood can appear in the fish surface of sick fish, and is attended by hemotoncus or white lump, i.e. scab.

The development of Along with computer technology and image processing techniques; Utilize image processing techniques can analyze to the scab of the fish surface of sick fish; And through analyzing the kind of confirming fish disease; And then sick fish isolated effectively or treat, thereby reduce the loss that the explosive diseases of aquaculture fish is brought effectively.

But after prior art is obtained the fish volume image of disease fish, can't be fast and accurately the scab of the fish surface of sick fish be extracted, to the collection of scab with extract and still need to carry out cutting and carry out through the manual fish volume image of operator to sick fish.Carry out cutting through the manual fish volume image of operator, reduced speed and accuracy that scab extracts on the one hand sick fish.Improved operator's labour intensity on the other hand, and, also be unfavorable for the large-scale promotion in industry because the operator need possess the professional knowledge of computing machine and Flame Image Process aspect.

Therefore, how from the fish volume image of sick fish, to obtain the scab image fast and accurately, just become problem demanding prompt solution.

Summary of the invention

The present invention provides fish volume image disposal route and the system of a kind of sick fish.

The present invention provides the fish volume image disposal route of a kind of sick fish, comprising:

Obtain the fish volume image of disease fish;

Extract the normalization component of R, G and three colors of B of said fish volume image rgb space respectively;

Said normalization component to extracting carries out the medium filtering image enhancement processing respectively;

Obtain the picture element matrix that carries out the said fish volume image after the said medium filtering image enhancement processing;

According to grey scale pixel value the pixel in the said picture element matrix being carried out whole normalization handles;

Fish volume image to carrying out after said whole normalization is handled carries out the piecemeal processing;

Variance to each piecemeal in the said fish volume image after the piecemeal processing is carried out threshold decision, to obtain fish body region image;

From said fish body region image, obtain unusual target area image;

Remove the non-scab target area image in the said unusual target area image, to obtain the scab target image.

The present invention provides the fish volume image disposal system of a kind of sick fish, comprising:

Image capture module is used to obtain the fish volume image of disease fish;

The fish body is cut apart module, is used for extracting respectively the normalization component of R, G and three colors of B of said fish volume image rgb space; Said normalization component to extracting carries out the medium filtering image enhancement processing respectively; Obtain the picture element matrix that carries out the said fish volume image after the said medium filtering image enhancement processing; According to grey scale pixel value the pixel in the said picture element matrix being carried out whole normalization handles; Fish volume image to carrying out after said whole normalization is handled carries out the piecemeal processing; Variance to each piecemeal in the said fish volume image after the piecemeal processing is carried out threshold decision, to obtain fish body region image;

Unusual target area image acquisition module is used for obtaining unusual target area image from said fish body region image;

Scab target image acquisition module is used for removing the non-scab target area image of said unusual target area image, to obtain the scab target image.

Fish volume image disposal route of sick fish of the present invention and system can be accurately and obtain the corresponding scab target image of the scab of disease fish easily; For the automatic fish disease identification of computing machine provides basic data; Thereby be convenient to the technician in time understand the disease fish the concrete condition of ill disease; In time take measures to contain the spreading of explosive disease of aquaculture fish; Thereby the loss that the explosive disease that reduces the aquaculture fish is brought, and the fish volume image disposal route of the sick fish of invention is easy and simple to handle, has good adaptability.

Description of drawings

In order to be illustrated more clearly in the embodiment of the invention or technical scheme of the prior art; To do one to the accompanying drawing of required use in embodiment or the description of the Prior Art below introduces simply; Obviously, the accompanying drawing in describing below is some embodiments of the present invention, for those of ordinary skills; Under the prerequisite of not paying creative work property, can also obtain other accompanying drawing according to these accompanying drawings.

Fig. 1 is the process flow diagram of the fish volume image disposal route embodiment one of the sick fish of the present invention;

Fig. 2 is the process flow diagram of the fish volume image disposal route embodiment two of the sick fish of the present invention;

Fig. 3 is the process flow diagram of the fish volume image disposal route embodiment three of the sick fish of the present invention;

Fig. 4 is the process flow diagram of the fish volume image disposal route embodiment four of the sick fish of the present invention;

Fig. 5 is the structural drawing of the fish volume image disposal system embodiment one of the sick fish of the present invention.

Embodiment

For the purpose, technical scheme and the advantage that make the embodiment of the invention clearer; To combine the accompanying drawing in the embodiment of the invention below; Technical scheme in the embodiment of the invention is carried out clear, intactly description; Obviously, described embodiment is the present invention's part embodiment, rather than whole embodiment.Based on the embodiment among the present invention, those of ordinary skills are not making the every other embodiment that is obtained under the creative work prerequisite, all belong to the scope of the present invention's protection.

Fig. 1 is the process flow diagram of the fish volume image disposal route embodiment one of the sick fish of the present invention, and as shown in Figure 1, the method for present embodiment can may further comprise the steps:

Step 101, obtain the fish volume image of disease fish.

For instance, the fish volume image of sick fish can through charge coupled cell under water (Charge-coupledDevice, hereinafter to be referred as: CCD) camera and coupled image pick-up card obtain in real time.Need to prove; How to distinguish fish volume image of disease fish and the fish volume image of normal fish and do not belong to category of the present invention; Those skilled in the art can make a distinction the fish volume image of the sick fish fish volume image with normal fish through prior art or artificial means, thereby obtain carrying out the fish volume image of the sick fish that the fish volume image of disease fish handles.

Step 102, from the fish volume image, obtain fish body region image.

For instance; After obtaining the fish volume image of disease fish; Can carry out adaptive threshold through the fish volume image to the sick fish that obtains cuts apart; And then from the fish volume image of the sick fish that comprises background, extract fish body region image, thus can reduce the data operation quantity of successive image treatment step effectively, improve the treatment effeciency of the fish volume image of disease fish.

Step 103, from fish body region image, obtain unusual target area image.

For instance, after obtaining the fish body region image of sick fish, can further obtain the unusual target area image of the scab position in the fish body region of disease fish through step 102.So just can the scab of sick fish be split from the fish body region of sick fish, thereby be convenient to the carrying out of subsequent step, can further reduce the data operation quantity of successive image treatment step effectively, thereby improve the treatment effeciency of the fish volume image of disease fish.

Step 104, remove the non-scab target area image in the unusual target area image, to obtain the scab target image.

For instance; After obtaining unusual target area image through step 103 again; For reducing the identification accuracy for processing of scab, reduce the data volume of the identification processing of scab simultaneously, can further remove the normal part of the fish body in the unusual target area image; After the non-scab target area image in removing unusual target area image, just can obtain only to comprise the scab target image of scab.After obtaining the scab target image; Through the scab target image is carried out feature extraction just can understand the disease fish the relevant information such as kind, the order of severity of ill disease; The professional of agriculture just can take appropriate measures; Means such as isolation, drug treatment for example, thereby the spreading of explosive disease of in time containing the aquaculture fish reduce the loss that the explosive diseases of aquaculture fish is brought.Need to prove; The present invention does not relate to through the scab target image is carried out feature extraction; Understand the disease fish the concrete mode of relevant information such as kind, the order of severity of ill disease; Those skilled in the art can adopt the image processing method of prior art to realize the processing of scab target image is repeated no more here.

The fish volume image disposal route of the sick fish of present embodiment can be accurately and is obtained the corresponding scab target image of the scab of disease fish easily; For the automatic fish disease identification of computing machine provides basic data; Thereby be convenient to the technician in time understand the disease fish the concrete condition of ill disease; In time take measures to contain the spreading of explosive disease of aquaculture fish; Thereby the loss that the explosive disease that reduces the aquaculture fish is brought, and the fish volume image disposal route of the sick fish of present embodiment is easy and simple to handle, has good adaptability.

Fig. 2 is the process flow diagram of the fish volume image disposal route embodiment two of the sick fish of the present invention, and as shown in Figure 2, present embodiment is the further refinement to embodiment one, and the method for present embodiment can may further comprise the steps:

Step 201, obtain the fish volume image of disease fish.

Step 202, extract the normalization component of R, G and three colors of B of fish volume image rgb space respectively.

For instance; Obtain the fish volume image of sick fish through step 201 after; Can extract processing to R (Red), G (Green) and three colors of B (Blue) of the rgb space of this fish volume image, thereby obtain the normalization component of R, G and three colors of B, handle for subsequent step.

Preferably, the normalization of extracting R, G and three colors of B of fish volume image rgb space respectively divide measuring can for: obtain the normalization component of fish volume image rgb space through formula (1),

r＝R/(R+B+G)

g＝G/(R+B+G) (1)

b＝B/(R+G+B)

Wherein, r, g and b are respectively the normalization component of R, G and three colors of B of said fish volume image rgb space.

Step 203, the normalization component that extracts is carried out the medium filtering image enhancement processing respectively.

For instance; Extract the normalization component of R, G and three colors of B of fish volume image rgb space through step 202 after; Can carry out the medium filtering image enhancement processing to the normalization component of R, G and three colors of B extracted, thereby be convenient to the carrying out of subsequent step.Normalization component to R, G and three colors of B carries out the acutance that the medium filtering image enhancement processing can improve the fish volume image, thereby accuracy and the precision of obtaining fish body region image are had significant improvement.

Preferably, to the normalization component that extracts carry out respectively the medium filtering image enhancement processing specifically can for: through template pixel scanning fish volume image, the center of template pixel overlaps with the target pixel location of said fish volume image; The gray-scale value of the pixel of the fish volume image in the calculating pixel template; The pixel of fish volume image is arranged by the gray-scale value size again, to obtain pixel median; With the pixel value of pixel median as object pixel.

For instance; Template pixel can be that size is the template pixel of 3 pixels * 3 pixels; Template pixel is roamed in the fish volume image, thereby reached the purpose of scanning fish volume image, template pixel specifically can pass through line by line or pursue each pixel of the form scanning fish volume image of row.In scanning process; The center of template pixel overlaps with certain locations of pixels of fish volume image; When the center of template pixel overlaps with certain locations of pixels of fish volume image; The pixel of the fish volume image that soon overlaps with the center of template pixel is as the object pixel of fish volume image, thereby reaches the purpose that each pixel to the fish volume image all scans.

After each confirmed object pixel, the gray-scale value of each pixel of the fish volume image in the calculating pixel template was arranged the pixel of fish volume image by the gray-scale value size, to obtain pixel median again.For example for as the pixel a of the fish volume image of object pixel (i, j), when template pixel was the template pixel of pixel * 3 pixels, the fish volume image in the template pixel had 9 pixels, these 9 pixels are pressed gray-scale value x _iBe respectively (x ₁, x ₂, x ₃, x ₄, x ₅, x ₆, x ₇, x ₈, x ₉), with these 9 gray-scale value (x ₁, x ₂, x ₃, x ₄, x ₅, x ₆, x ₇, x ₈, x ₉) descendingly arrange again, and the gray-scale value that will be positioned at the centre position is as pixel median.Promptly when template pixel is made up of n pixel; If n is an odd number; Then

is pixel median; If n is an even number, then

is pixel median.After confirming pixel median, just can be with the pixel value of pixel median as object pixel.After each pixel of fish volume image all carried out above-mentioned processing, just can make the gray-scale value of each pixel of fish volume image all equal the pixel median of this pixel, thereby can make the fish volume image obtain sharpening effectively, be convenient to the processing of subsequent step as object pixel.

Step 204, carry out adaptive threshold and cut apart carrying out fish volume image after the medium filtering image enhancement processing, to obtain fish body region image.

For instance; After step 203 is obtained to carry out the fish volume image of medium filtering image enhancement processing; Can carry out the adaptive threshold dividing processing to this fish volume image, remove the background parts of fish volume image, and then obtain the fish body region image of the fish body region of the sick fish of performance.

Carry out adaptive threshold and cut apart carrying out fish volume image after the medium filtering image enhancement processing, specifically can for: obtain the picture element matrix that carries out the fish volume image after the medium filtering image enhancement processing; According to grey scale pixel value the pixel in the picture element matrix being carried out whole normalization handles; Fish volume image to carrying out after whole normalization is handled carries out the piecemeal processing; Fish volume image to after the piecemeal processing carries out threshold decision, to obtain fish body region image.

For instance, obtain the picture element matrix that carries out the fish volume image after the medium filtering image enhancement processing, for example; To carry out gray scale fish volume image I after the medium filtering image enhancement processing and be defined as the picture element matrix of XY; Wherein I (i, j) i is capable in the representing matrix, the grey scale pixel value of the fish volume image of j row; I1 is the picture element matrix of N * N, and I1 is a block of image I.

According to grey scale pixel value the pixel in the picture element matrix is carried out whole normalization and handles, specifically can for: definition Mean (I) be the whole average of fish volume image, and Var (I) is whole variance, and then the normalization processing can be passed through formula (2) to (4) and realizes.

$\mathrm{Mean}\left(I\right)=1/\mathrm{XY}\underset{i=0}{\overset{X}{\mathrm{\Σ}}}\underset{j=0}{\overset{Y}{\mathrm{\Σ}}}I(i,j)---\left(3\right)$

$\mathrm{Var}\left(I\right)=\mathrm{XY}\underset{i=0}{\overset{X}{\mathrm{\Σ}}}\underset{j=0}{\overset{Y}{\mathrm{\Σ}}}I(i,j)-\mathrm{Mean}{\left(I\right)}^2---\left(4\right)$

Wherein, (i is j) for carrying out the fish volume image gray values of pixel points after normalization is handled for G; M ₀And Var ₀Be respectively the average and the variance of the mathematical expectation of respective pixel gray-scale value.

Fish volume image to carrying out after whole normalization is handled carries out the piecemeal processing; Specifically can for: get the piecemeal that is of a size of N pixel * N pixel; The fish volume image that will carry out after whole normalization is handled is divided into equal-sized some, can represent the average of the piecemeal of fish volume image with M (I1), the variance of the piecemeal of Var (I1) expression fish volume image; And then through formula (5) and (6) calculating (i j) is the average and the variance of the piecemeal at center with pixel I.

$M\left(I1\right)=1/(N\×N)\underset{i=0}{\overset{N}{\mathrm{\Σ}}}\underset{j=0}{\overset{N}{\mathrm{\Σ}}}I(i,j)---\left(5\right)$

$\mathrm{Var}\left(\mathrm{Ir}\right)=1/(N\×N)\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}\underset{j=0}{\overset{N-1}{\mathrm{\Σ}}}{(I(i,j)-M\left(I1\right))}^2---\left(6\right)$

Fish volume image after piecemeal handled carries out threshold decision, to obtain fish body region image, specifically can for: each piecemeal is judged, to confirm that corresponding piecemeal is the transitional region between background area, foreground area or background area and the foreground area.For example, as the variance Var of certain piecemeal I1 (I1) during, can confirm directly that this piecemeal belongs to background area less than threshold value T1; If it is isolated point or noise spot place piecemeal that the variance of this piecemeal between [T1, T2], then can be confirmed this piecemeal; If the variance of piecemeal between [T2, T3], can confirm then that this piecemeal is a transitional region; Except that above situation, the piecemeal that possesses the variance in other numerical ranges can be confirmed as foreground area.Need to prove, since varying environment, the different light rays condition; Different types of sick fish; Its threshold value T1, T2 and T3 that is used to carry out that piecemeal handles is different, and therefore, those skilled in the art need be provided with voluntarily according to actual needs threshold value T1, T2 and T3.

When definite a certain piecemeal is transitional region, further need (i j), calculates the gray average M (I1) with the pixel in the zone of the centrical N pixel of this pixel * N pixel through formula (5) to the pixel I in this piecemeal.If this gray values of pixel points greater than this average, is then further confirmed as the background pixel point with this pixel, otherwise this pixel is confirmed as the foreground pixel point.

Through keeping the piecemeal and the foreground pixel point of foreground area, just can the background parts in the fish volume image be rejected, thereby obtain the corresponding fish body region image that reflects the fish body region of sick fish.

Step 205, from fish body region image, obtain unusual target area image.

Step 206, remove the non-scab target area image in the unusual target area image, to obtain the scab target image.

Because step 201, step 205 is corresponding with step 206 with the realization principle and the technique effect of step 101, step 103 and the step 104 of embodiment one similar, repeat no more here.

The fish volume image disposal route of the sick fish of present embodiment can accurately and easily be rejected the background parts in the fish volume image of sick fish; Thereby obtain the fish body region image of the fish body region of the sick fish of corresponding reflection; The corresponding accuracy that can improve the corresponding scab target image of the scab that obtains the disease fish; Thereby be convenient to the technician in time understand the disease fish the concrete condition of ill disease; Thereby can in time take measures to contain the spreading of explosive disease of aquaculture fish, reduce the loss that the explosive diseases of aquaculture fish is brought.

Fig. 3 is the process flow diagram of the fish volume image disposal route embodiment three of the sick fish of the present invention, and as shown in Figure 3, present embodiment is the further refinement to embodiment one and two, and the method for present embodiment can may further comprise the steps:

Step 301, step are obtained the fish volume image of disease fish.

Step 302, from the fish volume image, obtain fish body region image.

Step 303, obtain the gray-scale value of the pixel of fish body region image.

For instance, specifically can be through fish body region image be defined as M pixel * N pixel, f (m, n) be corresponding fish body region pattern matrix in the position (m; N) grey scale pixel value of locating, K are the corresponding number of greyscale levels of fish body region image, and f (m, n) ∈ { 0; 1 ..., K-1}.

Step 304, obtain the one dimension gray-scale statistical histogram data of fish body region image.

For instance, the one dimension gray-scale statistical histogram data of fish body region image can be through being defined as H (k) histogram functions of fish body region image, and calculate corresponding histogram data through formula (7).

$H\left(k\right)=\underset{m=0}{\overset{M-1}{\mathrm{\Σ}}}\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}\mathrm{\δ}[f(m,n)-k],$ k∈{0，1，...，K-1} (7)

Wherein, function δ (0)=1, δ (k ≠ 0)=0, H (k) is for having the number of pixels of gray level k.

Step 305, according to cluster classification number and iteration stopping threshold value, calculate to divide matrix through the cluster objective function, to obtain unusual target area image.

For instance, behind step 304 acquisition histogram data, can pass through cluster objective function J _m(U V:K) calculates the division matrix, specifically can pass through formula (8) and realize.

$J_m(U,V:K)=\underset{i=1}{\overset{c}{\mathrm{\Σ}}}\underset{k=0}{\overset{K-1}{\mathrm{\Σ}}}{\left(u_{\mathrm{ik}}\right)}^{m}H\left(k\right)\mathrm{Pk}-v_i{P}_A^2---\left(8\right)$

Wherein, formula (8) satisfies extremum conditions:

K ∈ 0,1 ..., K-1}, m and c are cluster classification number, and m＞1,2≤c≤K-1; u _IkBe the degree of membership of i class sample k, v _iBe cluster centre, k-v _iExpression sample point k is apart from cluster centre v _iEuclidean distance, and k ∈ R ^s, v _i∈ R ^s, s is the dimension in cluster space; U={u _IkThe matrix of (K-1) * c of expression dimension, V={v ₁, v ₂..., v _cThe expression a s * c matrix; K is the number of greyscale levels of fish body region image.

After confirming the cluster objective function, need given cluster classification to count c (2≤c≤K-1) and m (m＞1), and setting iteration stopping value ε=0.001; Initialization cluster prototype pattern V ⁽⁰⁾, and set iteration count b=0.

And then through formula (9) calculating division matrix U ^(b)

$u_{\mathrm{ik}}=\left|\underset{j=1}{\overset{c}{\mathrm{\Σ}}}\right|\frac{\mathrm{Pl}-v_i{P}_A}{\mathrm{Pl}-v_j{P}_A}{|}^{2/(m-1)}{|}^{-1}---\left(9\right)$

And through formula (10) calculating cluster prototype pattern V ^(b+1)

$V_i^{(b+1)}=\frac{\underset{k=0}{\overset{K-1}{\mathrm{\Σ}}}{\left(u_{\mathrm{ik}}\right)}^{m}H\left(k\right)\mathrm{gk}}{\underset{k=0}{\overset{K-1}{\mathrm{\Σ}}}{\left(u_{\mathrm{ik}}\right)}^{m}H\left(k\right)}---\left(10\right)$

If

Then stop to calculate, promptly obtain dividing matrix U ^(b), otherwise make b=b+1, and recomputate the division matrix U ^(b)With cluster prototype pattern V ^(b+1), up to obtaining satisfactory division matrix U ^(b)So just can unusual target area image be divided from fish body region image, for the processing of subsequent step.

Step 306, remove the non-scab target area image in the unusual target area image, to obtain the scab target image.

Because step 301, step 302 is corresponding with step 306 with the realization principle and the technique effect of step 101, step 102 and the step 104 of embodiment one similar, repeat no more here.

The fish volume image disposal route of the sick fish of present embodiment can accurately and easily divide out with the unusual target area image of the scab in the fish body region image of sick fish; The corresponding accuracy that can improve the corresponding scab target image of the scab that obtains the disease fish; Thereby be convenient to the technician in time understand the disease fish the concrete condition of ill disease; Thereby can in time take measures to contain the spreading of explosive disease of aquaculture fish, reduce the loss that the explosive diseases of aquaculture fish is brought.

Fig. 4 is the process flow diagram of the fish volume image disposal route embodiment four of the sick fish of the present invention, and as shown in Figure 4, present embodiment is the further refinement to embodiment one and two, and the method for present embodiment can may further comprise the steps:

Step 401, step are obtained the fish volume image of disease fish.

Step 402, from the fish volume image, obtain fish body region image.

Step 403, from fish body region image, obtain unusual target area image.

Because step 401 to step 403 is corresponding and the realization principle and the technique effect of step 101 to the step 103 of embodiment one are similar, repeats no more here.

Step 404, unusual target area image is reversed.

For instance; Owing to including non-scab image information in the unusual target area image of obtaining through step 403; In order effectively non-scab image information to be removed from unusual target area image; Thereby obtain accurate scab image information, need be through the unusual target area image that includes non-scab information that step 404 pair is obtained through step 403 processing of overturning.

Step 405, from fish body region image, obtain normal target area image piece, with normal target area image piece as scan templates.

For instance, can obtain the normal target area image piece of normal part of the fish body of the sick fish of performance from the optional position of fish body region image, and with this normal target area image piece as scan templates.Need to prove that normal target area image piece should keep certain distance with unusual target area image as far as possible, does not comprise any image information relevant with scab thereby make in the normal target area image piece.

The reduction of reversing of step 406, the unusual target area image after will reversing.

For instance, behind step 405 acquisition scan templates, can make it present original location status the reduction of reversing through the unusual target area image of step 404 counter-rotating, handle for subsequent step.

Step 407, use scan templates scan unusual target area image.

For instance, can use the mode with the preferred embodiment of embodiment two to scan unusual target area image through scan templates, short and sweet for making instructions, concrete operations step and flow process repeat no more.

Step 408, the eigenwert of calculating the unusual target area image under the scan templates and the average of computation of characteristic values.

For instance, using scan templates to scan in the process of unusual target area image, the eigenwert of whole pixels of the unusual target area image under can the writing scan template, and the pairing average C of eigenwert ₁

Step 409, obtain the scab target image according to the average and the scanning threshold value of the eigenwert of scan templates.

For instance, the average of the eigenwert that the eigenwert of scan templates and average can be through asking for whole pixels that corresponding normal target area image piece comprised, and with the average C of this average as the eigenwert of scan templates ₂, and set the scanning threshold epsilon.Like this according to the average C of the eigenwert of scanning threshold epsilon and scan templates ₂Just can be to the corresponding pairing average C of eigenwert of each pixel of target area image unusually ₁Judge, when | C ₁-C ₂| during≤ε, then the pixel value with the respective pixel of the unusual target area image under the scan templates is made as 1, otherwise keeps it constant, thereby can the non-scab image information in the unusual target area image be removed, and then obtains corresponding scab target image.

The fish volume image disposal route of the sick fish of present embodiment can be accurately and is obtained the corresponding scab target image of the scab of disease fish easily; Thereby be convenient to the technician in time understand the disease fish the concrete condition of ill disease; Thereby can in time take measures to contain the spreading of explosive disease of aquaculture fish; The loss that the explosive disease of reduction aquaculture fish is brought, and the fish volume image disposal route of the sick fish of present embodiment is easy and simple to handle, has good adaptability.

Fig. 5 is the structural drawing of the fish volume image disposal system embodiment one of the sick fish of the present invention, and as shown in Figure 5, the fish volume image disposal system of the sick fish of present embodiment can comprise: image capture module 1 is used to obtain the fish volume image of disease fish; The fish body is cut apart module 2, is used for obtaining fish body region image from the fish volume image; Unusual target area image acquisition module 3 is used for obtaining unusual target area image from fish body region image; Scab target image acquisition module 4 is used for removing the non-scab target area image of unusual target area image, to obtain the scab target image.

The fish volume image disposal system of the sick fish of present embodiment can be used to carry out method embodiment illustrated in fig. 1, and its realization principle and technique effect are similar, repeat no more here.

One of ordinary skill in the art will appreciate that: all or part of step that realizes said method embodiment can be accomplished through the relevant hardware of programmed instruction; Aforesaid program can be stored in the computer read/write memory medium; This program the step that comprises said method embodiment when carrying out; And aforesaid storage medium comprises: various media that can be program code stored such as ROM, RAM, magnetic disc or CD.

What should explain at last is: above embodiment is only in order to explaining technical scheme of the present invention, but not to its restriction; Although with reference to previous embodiment the present invention has been carried out detailed explanation, those of ordinary skill in the art is to be understood that: it still can be made amendment to the technical scheme that aforementioned each embodiment put down in writing, and perhaps part technical characterictic wherein is equal to replacement; And these are revised or replacement, do not make the spirit and the scope of the essence disengaging various embodiments of the present invention technical scheme of relevant art scheme.

Claims (6)

1. the fish volume image disposal route of a sick fish is characterized in that, comprising:

Obtain the fish volume image of disease fish;

Extract the normalization component of R, G and three colors of B of said fish volume image rgb space respectively;

Said normalization component to extracting carries out the medium filtering image enhancement processing respectively;

Obtain the picture element matrix that carries out the said fish volume image after the said medium filtering image enhancement processing;

According to grey scale pixel value the pixel in the said picture element matrix being carried out whole normalization handles;

Fish volume image to carrying out after said whole normalization is handled carries out the piecemeal processing;

Variance to each piecemeal in the said fish volume image after the piecemeal processing is carried out threshold decision, to obtain fish body region image;

From said fish body region image, obtain unusual target area image;

Remove the non-scab target area image in the said unusual target area image, to obtain the scab target image.

2. method according to claim 1 is characterized in that, the said normalization component that extracts R, G and three colors of B of said fish volume image rgb space respectively is specially:

Through the normalization component of the said fish volume image rgb space of formula (1) acquisition,

r=R/（R+B+G）

g=G/（R+B+G） （1）

b=B/（R+G+B）

Wherein, r, g and b are respectively the normalization component of R, G and three colors of B of said fish volume image rgb space.

3. method according to claim 1 is characterized in that, saidly the said normalization component that extracts is carried out the medium filtering image enhancement processing respectively is specially:

Scan said fish volume image through template pixel, the center of said template pixel overlaps with the target pixel location of said fish volume image;

Calculate the gray-scale value of the pixel of the said fish volume image in the said template pixel;

The pixel of said fish volume image is arranged by the gray-scale value size again, to obtain pixel median;

With the pixel value of said pixel median as object pixel.

4. according to the described method of arbitrary claim in the claim 1 to 3, it is characterized in that, saidly from said fish body region image, obtain unusual target area image and be specially:

Obtain the gray-scale value of the pixel of said fish body region image;

Obtain the one dimension gray-scale statistical histogram data of said fish body region image;

According to cluster classification number and iteration stopping threshold value, calculate the division matrix through the cluster objective function, to obtain said unusual target area image.

5. according to the described method of arbitrary claim in the claim 1 to 3, it is characterized in that the said non-scab target area image that removes in the said unusual target area image is specially:

Said unusual target area image is reversed;

From said fish body region image, obtain normal target area image piece, with said normal target area image piece as scan templates;

With the said unusual target area image reduction of reversing after the counter-rotating;

Use said scan templates to scan said unusual target area image;

Calculate the eigenwert of the said unusual target area image under the said scan templates and calculate the average of said eigenwert;

Average according to the eigenwert of said scan templates obtains said scab target image with the scanning threshold value.

6. the fish volume image disposal system of a sick fish is characterized in that, comprising:

Image capture module is used to obtain the fish volume image of disease fish;

The fish body is cut apart module, is used for extracting respectively the normalization component of R, G and three colors of B of said fish volume image rgb space; Said normalization component to extracting carries out the medium filtering image enhancement processing respectively; Obtain the picture element matrix that carries out the said fish volume image after the said medium filtering image enhancement processing; According to grey scale pixel value the pixel in the said picture element matrix being carried out whole normalization handles; Fish volume image to carrying out after said whole normalization is handled carries out the piecemeal processing; Variance to each piecemeal in the said fish volume image after the piecemeal processing is carried out threshold decision, to obtain fish body region image;

Unusual target area image acquisition module is used for obtaining unusual target from said fish body region image.

Images